Impact of Fire on Herbivores

Fire plays an important role in the ecology and evolution of savanna and grassland ecosystems and is often used as a tool for the management of grassland vegetation. Mammalian herbivores have been shown to be attracted to recently burned areas. This attraction has largely been attributed to increases in plant nutrient content. However, because burning causes a reduction in vegetation height and increases in sighting distances, these areas may also be safer habitats from predators. To test these two different hypotheses we measured herbivore abundances as well as vegetation traits on unburned and newly burned areas in Serengeti National Park, Tanzania. The results of this study showed that regardless of body size herbivores prefer burned areas. Both vegetation nutrient quality and vegetation height (a proxy for predation pressure) play a role in explaining this preference for burned areas. However, the preference for burned areas and positive changes in vegetation characteristics is short term and lasts less than six months after burning. Despite short term increases in nutrient quality, increases in fire frequency (the number of times a site burns) cause changes in plant species composition towards species of lower nutrient quality. 

Currently, I am trying to determine how fire frequency impacts herbivore habitat preference and whether preference is the same in both North American and South African savanna grassland ecosystems. In order to test this I am using long-term data sets on herbivore distributions in sites that are burned annually, intermediately (3-4 year fire return interval) or unburned (last burned at least 20 years ago). These data sets have been collected at the Konza long-term ecological research site in Kansas and at Kruger National Park, South Africa.      


Impact of Fire on Carnivores

Currently, little is known about how fire affects short-term carnivore distributions across the landscape. Through the use of a long-term data set collected by the Serengeti Lion Project, we investigated the distribution of lions in relation to burned areas in Serengeti National Park.  We found that lions avoid burned areas despite the increases in prey availability in these areas. This avoidance may be due to the above mentioned reduction in vegetation height seen in burned areas which may cause decreased hunting success.

Impact of Fire and Grazing on Plants

Historically, fire played a large role in savanna and grassland ecosystems, as both lightning and human set fires frequently occurred. Humans are altering fire regimes by increasing/reducing the frequency of fires or completely excluding fires. Grazing also played a large role historically in these ecosystems, as many were home to large numbers of native megafauna. Currently, humans are altering grazing by completely removing native grazers or replacing them with livestock. It is important to have a more general understanding of the impacts of these drivers on these ecosystems in order to predict how they will behave with continued human influence. It is especially important to determine whether these drivers behave the same in systems with different evolutionary histories, biome ages, soil fertility and megaherbivore diversities.

Studies have suggested that fire and grazing influence South African (SA) and North American (NA) savanna grasslands in fundamentally different ways. Yet no methodologically identical studies have been done concurrently in SA and NA to allow for direct comparisons of results. Until, seven years ago study sites were established in Kruger National Park, SA and the Konza prairie long-term ecological research site in Kansas, NA. In order to test the impacts of fire and grazing on plant species composition (richness, diversity, evenness) exclosures and corresponding grazed paired plots were established on fire treatments that burn annually or never burn.

There is a large difference in herbivore diversity between the two sites (one species at Konza and 14 at Kruger). In an attempt to remove any effect that the greater herbivore diversity at Kruger might have, sites were established inside a buffalo enclosure. This allowed us to compare the impact of a single grazer at Kruger (Cape buffalo, Syncerus caffer) to a single grazer at Konza (Bison, Bos bison). We found a decrease in grass and forb richness, evenness and diversity over time when a single grazer was removed from each site. However, in Kruger these changes only occurred with burning. At both sites these changes occurred because of increases in abundance of dominant grass species.  

Ecosystem State Changes from Grasslands to Woodlands

Some studies have shown that when ecosystems change from one state to another there may be leading indicators as the ecosystem approaches the critical transition point. Studies have indicated that there may be rising standard deviation, elevated spatial variability, and increases in the coefficient of variation, skewness, autocorrelation, and spatial correlation. The ability to predict ecosystem state changes prior to their occurrence would be beneficial to park and ecosystem managers throughout the world. If changes in ecosystems were predicted in advance this might allow for management practices to be altered prior to the shift preventing it from happening. One way to empirically validate leading indicators is to compute spatial indicators, prior to and through the time when a regime shift is occurring. However, due to the lack of temporal data, testing for leading indicators has rarely occurred. To address this we are using a 'space-for-time-substitution' approach by looking at transitions in regime shifts across an ecosystem. Specifically we are looking for leading indicators of transitions from grassland to woodland states in the Serengeti-Mara ecosystem in Tanzania and Kenya, East Africa.